Abstract
This paper presents a benchmark for peak detection algorithms employed in fiber Bragg grating spectrometric interrogation systems. The accuracy, precision, and computational performance of currently used algorithms and those of a new proposed artificial neural network algorithm are compared. Centroid and gaussian fitting algorithms are shown to have the highest precision but produce systematic errors that depend on the FBG refractive index modulation profile. The proposed neural network displays relatively good precision with reduced systematic errors and improved computational performance when compared to other networks. Additionally, suitable algorithms may be chosen with the general guidelines presented.
Highlights
Fiber Bragg grating (FBG) interrogation techniques form a mature field of research where computational techniques must be used to improve the process of monitoring FBG sensors
The values for the different neural network training algorithms are shown in Table 2, where the resulting mean mean square error (MSE) and mean epochs needed to reach any of the stopping conditions are presented
With respect to the FBG signal corresponding to the sensor with uniform modulation profile, the results presented in Figures 9 and 10 support the conclusion that the gaussian fitting and the centroid are more accurate and precise than the other algorithms, and have a relatively good noise tolerance
Summary
Fiber Bragg grating (FBG) interrogation techniques form a mature field of research where computational techniques must be used to improve the process of monitoring FBG sensors. Many techniques which are part of commercial systems use a periodically tunable laser source to illuminate the FBG and produce the signal corresponding to the spectrum of the light that interacts with the device [1], characterizing the spectrometric technique. These wavelength sweeping techniques may have high accuracy and precision, requiring an additional computational processing of the acquired signals, and a wavelength signal reference which is systematically used during the sensor interrogation process [2]. The approximation occurs during the training phase of the neural network and is applied whenever the peak identification is necessary [3,4]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
